感性理解 EOJ 测评机速度 - Blog

Compare

ultmaster edited 6 年，8 月前

This is a past version of blog 感性理解 EOJ 测评机速度

膜 LOJ。以及抄袭原文。

对比测试

#	Language	Loop	Euler Sieve	Floyd-Warshall	`std::set`	Memory Alloc (new)
Codeforces	G++ 5.1.0 (-O2)	288	295	967	701	857
LibreOJ(2017.12.10)	G++ 5.4.0 (-O2)	301	319	823	850	736
Universal Online Judge	G++ 4.8.4 (-O2)	435	530	1377	996	918
BZOJ	G++ 4.4.5 (-O2)	504	1280	2092	1368	1568
HUSTOJ	G++ 4.4.5 (-O2)	908	1460	1413	1553	Unexpected MLE (ML = 512MB)
EOJ (Goat)	G++ 5.4.0 (-O2)	464	344	1120	672	736 (ML = 1GB)
EOJ (Cow)	G++ 5.4.0 (-O2)	332	388	1056	736	768 (ML = 1GB)

It seems that EOJ is a little bit slower than Codeforces, but fast enough.

测试代码

// Test 1: Loop
#include<cstdio>

using namespace std;

int main(){
    int a=1000000000,b=1;
    while(a)b<<=1,a--;
    printf("%d\n",b);
    return 0;
}

// Test 2: Euler Sieve
#include<cstdio>

using namespace std;
const int MX=50000000;
int p[MX],m[MX],pc;
int main(){
    for(int i=2;i<MX;i++){
        if(!m[i])p[++pc]=m[i]=i;
        static int k;
        for(int j=1;j<=pc&&p[j]<=m[i]&&(k=p[j]*i)<MX;j++)m[k]=p[j];
    }
    int ans=0;
    for(int i=1;i<=pc;i++)ans^=p[i];
    printf("%d\n",ans);
    return 0;
}

// Test 3: Floyd-Warshall
#include<cstdio>

using namespace std;
const int MX=1000;
int G[MX][MX];
int sed=0;
inline int rand(){return sed=(sed*sed*73+sed*233+19260817)&0x0000ffff;}
int main(){
    for(int i=0;i<MX;i++)
        for(int j=0;j<MX;j++)
            G[i][j]=rand();
    for(int i=0;i<MX;i++)
        for(int j=0;j<MX;j++)
            for(int k=0;k<MX;k++)
                if(G[j][k]>G[j][i]+G[i][k])G[j][k]=G[j][i]+G[i][k];
    int ans=0;
    for(int i=0;i<MX;i++)
        for(int j=0;j<MX;j++)
            ans^=G[i][j];
    printf("%d\n",ans);
    return 0;
}

// Test 4: std::set
#include<cstdio>
#include<algorithm>
#include<set>

using namespace std;

const int MX=1000000;
int sed=0;
inline int rand(){return sed=(sed*sed*73+sed*233+19260817);}
int main(){
    set<int>S;
    for(int i=0;i<MX;i++)S.insert(rand());
    int ans=0;
    for(set<int>::iterator it=S.begin();it!=S.end();it++)ans^=*it;
    printf("%d\n",ans);
    return 0;
}

// Test 5: Memory Alloc (new)
#include<cstdio>

using namespace std;
const int MX=20000000;
int *it[MX];
int main(){
    for(int i=0;i<MX;i++)it[i]=new int;
    for(int i=0;i<MX;i++)*it[i]=i;
    int ans=0;
    for(int i=0;i<MX;i++)ans^=*it[i];
    printf("%d\n",ans);
    return 0;
}

IO test

3 million non-negative integers in one single line.

Input will be faster if it is naturally buffered (for example in Python).

Time limit is 5 seconds. All numbers are in millisecond(s).

#	Machine	$[0,2)$	$[0,8)$	$[0,2^{15})$	$[0,2^{31})$	$[0,2^{63})$
`fread`	Goat	20	12	52	72	120
`fread`	Cow	16	12	48	88	104
`getchar`	Goat	64	52	140	252	452
`getchar`	Cow	60	68	160	268	520
`cin.tie(0)`	Goat	460	560	764	1344	2044
`cin.tie(0)`	Cow	464	696	948	1344	3096
`scanf`	Goat	288	252	444	600	844
`scanf`	Cow	256	260	572	516	1148
`cin`	Goat	516	608	848	1340	2100
`cin`	Cow	514	516	860	1444	2456
Java	Goat	2732	3016	3368	3928	4440
Java	Cow	2572	2904	3184	4396	TL
Java Buffered	Goat	780	900	996	916	1864
Java Buffered	Cow	652	620	836	1220	1472
Python	Goat	836	992	1052	1052	1072
Python	Cow	984	616	672	812	948

膜 LOJ。以及抄袭原文。

对比测试

#	Language	Loop	Euler Sieve	Floyd-Warshall	`std::set`	Memory Alloc (new)
Codeforces	G++ 5.1.0 (-O2)	288	295	967	701	857
LibreOJ(2017.12.10)	G++ 5.4.0 (-O2)	301	319	823	850	736
Universal Online Judge	G++ 4.8.4 (-O2)	435	530	1377	996	918
BZOJ	G++ 4.4.5 (-O2)	504	1280	2092	1368	1568
HUSTOJ	G++ 4.4.5 (-O2)	908	1460	1413	1553	Unexpected MLE (ML = 512MB)
EOJ (Goat)	G++ 5.4.0 (-O2)	464	344	1120	672	736 (ML = 1GB)
EOJ (Cow)	G++ 5.4.0 (-O2)	332	388	1056	736	768 (ML = 1GB)

It seems that EOJ is a little bit slower than Codeforces, but fast enough.

测试代码

// Test 1: Loop
#include<cstdio>

using namespace std;

int main(){
    int a=1000000000,b=1;
    while(a)b<<=1,a--;
    printf("%d\n",b);
    return 0;
}

// Test 2: Euler Sieve
#include<cstdio>

using namespace std;
const int MX=50000000;
int p[MX],m[MX],pc;
int main(){
    for(int i=2;i<MX;i++){
        if(!m[i])p[++pc]=m[i]=i;
        static int k;
        for(int j=1;j<=pc&&p[j]<=m[i]&&(k=p[j]*i)<MX;j++)m[k]=p[j];
    }
    int ans=0;
    for(int i=1;i<=pc;i++)ans^=p[i];
    printf("%d\n",ans);
    return 0;
}

// Test 3: Floyd-Warshall
#include<cstdio>

using namespace std;
const int MX=1000;
int G[MX][MX];
int sed=0;
inline int rand(){return sed=(sed*sed*73+sed*233+19260817)&0x0000ffff;}
int main(){
    for(int i=0;i<MX;i++)
        for(int j=0;j<MX;j++)
            G[i][j]=rand();
    for(int i=0;i<MX;i++)
        for(int j=0;j<MX;j++)
            for(int k=0;k<MX;k++)
                if(G[j][k]>G[j][i]+G[i][k])G[j][k]=G[j][i]+G[i][k];
    int ans=0;
    for(int i=0;i<MX;i++)
        for(int j=0;j<MX;j++)
            ans^=G[i][j];
    printf("%d\n",ans);
    return 0;
}

// Test 4: std::set
#include<cstdio>
#include<algorithm>
#include<set>

using namespace std;

const int MX=1000000;
int sed=0;
inline int rand(){return sed=(sed*sed*73+sed*233+19260817);}
int main(){
    set<int>S;
    for(int i=0;i<MX;i++)S.insert(rand());
    int ans=0;
    for(set<int>::iterator it=S.begin();it!=S.end();it++)ans^=*it;
    printf("%d\n",ans);
    return 0;
}

// Test 5: Memory Alloc (new)
#include<cstdio>

using namespace std;
const int MX=20000000;
int *it[MX];
int main(){
    for(int i=0;i<MX;i++)it[i]=new int;
    for(int i=0;i<MX;i++)*it[i]=i;
    int ans=0;
    for(int i=0;i<MX;i++)ans^=*it[i];
    printf("%d\n",ans);
    return 0;
}

IO test

3 million non-negative integers in one single line.

Input will be faster if it is naturally buffered (for example in Python).

Time limit is 5 seconds. All numbers are in millisecond(s).

#	Machine	$[0,2)$	$[0,8)$	$[0,2^{15})$	$[0,2^{31})$	$[0,2^{63})$
`fread`	Goat	20	12	52	72	120
`fread`	Cow	16	12	48	88	104
`getchar`	Goat	64	52	140	252	452
`getchar`	Cow	60	68	160	268	520
`cin.tie(0)`	Goat	460	560	764	1344	2044
`cin.tie(0)`	Cow	464	696	948	1344	3096
`scanf`	Goat	288	252	444	600	844
`scanf`	Cow	256	260	572	516	1148
`cin`	Goat	516	608	848	1340	2100
`cin`	Cow	514	516	860	1444	2456
Java	Goat	2732	3016	3368	3928	4440
Java	Cow	2572	2904	3184	4396	TL
Java Buffered	Goat	780	900	996	916	1864
Java Buffered	Cow	652	620	836	1220	1472
Python	Goat	836	992	1052	1052	1072
Python	Cow	984	616	672	812	948

感性理解 EOJ 测评机速度 (Past)

对比测试

测试代码

IO test

对比测试

测试代码

IO test

服务

开源

关于我们